Canadian mining updates: bug-fixes, more SSHRC matches, less IRSC fal…

…se-positives, more relation terms.

iis-wf/iis-wf-referenceextraction/src/main/resources/eu/dnetlib/iis/wf/referenceextraction/project/main_sqlite/oozie_app/lib/scripts/projects.sql

@@ -67,56 +67,57 @@ union all
 select jdict('documentId', docid, 'projectId', id, 'confidenceLevel', 0.8, 'textsnippet', textsnippet) as C1, docid, id, fundingclass1, grantid
 from (
     select docid, case when regexprmatches(".*(?:(?:CIHR|IRSC)|(?i)(?:canad(?:ian|a) institute(?:s)? health research|institut(?:(?:e)?(?:s)?)? recherche sant(?:é|e) canada)).*", prev||" "||middle||" "||next) then (select id from grants where fundingclass1 = 'CIHR')
-                       when regexprmatches(".*(?:(?:NSERC|CRSNG)|(?i)(?:nat(?:ural|ional) science(?:s)?(?:\sengineering(?:\sresearch)?|\sresearch)? co(?:u)?n(?:c|se)(?:i)?l|conseil(?:s)? recherche(?:s)? science(?:s)? naturel(?:les)?(?:\sg(?:e|é)nie)? canada)).*", prev||" "||middle||" "||next) then (select id from grants where fundingclass1 = 'NSERC')
-                       when regexprmatches(".*(?:(?:SSHRC|CRSH)|(?i)(?:social sciences humanities|conseil(?:s)? recherche(?:s)?(?:\ssciences humaines)? canada)).*", prev||" "||middle||" "||next) then (select id from grants where fundingclass1 = 'SSHRC')
+                       when regexprmatches(".*(?:(?:NSERC|CRSNG)|(?i)(?:nat(?:ural|ional) science(?:s)?(?:\sengineering(?:\sresearch)?|\sresearch) co(?:u)?n(?:c|se)(?:i)?l|conseil(?:s)? recherche(?:s)? science(?:s)? naturel(?:les)?(?:\sg(?:e|é)nie)? canada)).*", prev||" "||middle||" "||next) then (select id from grants where fundingclass1 = 'NSERC')
+                       when regexprmatches(".*(?:(?:SSHRC|CRSH|SSRCC)|(?i)(?:social science(?:s)?|conseil(?:s)? recherche(?:s)?(?:\ssciences humaines)? canada|humanities\sresearch)).*", prev||" "||middle||" "||next) then (select id from grants where fundingclass1 = 'SSHRC')
                        else 'canadian_unspecified_id'
                   end as id, "unidentified" as grantid, "Canadian" as fundingclass1, (prev||" "||middle||" "||next) as textsnippet
     from
         (setschema 'docid,prev,middle,next' select c1, textwindow2s(filterstopwords(keywords(c2)), 15,1,15, "^(?:(?:(?:CIHR|IRSC)|(?:NSERC|CRSNG)|(?:SSHRC|CRSH))|(?i)(?:co(?:(?:un(?:cil|sel))|(?:nseil(?:s)?))|canad(?:a|ian)))$") from pubs where c2 is not null)
     where
     ( /* Terms */
         /* Acronyms */
-        regexprmatches("^(?:(?:CIHR|IRSC)|(?:NSERC|CRSNG)|(?:SSHRC|CRSH))$", middle)
+        regexprmatches("^(?:CIHR|(?:NSERC|CRSNG)|(?:SSHRC|CRSH|SSRCC))$", middle)
+        or (
+            regexprmatches("^IRSC$", middle) /* This is the french acronym of CIHR. It also refers to some other organizations, so we search and exclude them. */
+            and not regexprmatches(".*(?:informal relationships social capital|interlocus sexual conflict|international (?:rosaceae|rosbreed) snp consortium|iranian seismological cent(?:er|re)).*", lower(prev||" "||next))
+        )
         or (/* Full-names */
-            (   /* Middle: "Council", "Counsel", "Conseil", "Conseils" --> NSERC/CRSNG, SSHRC/CRSH */
+            (   /* Middle: "Council", "Counsel", "Conseil", "Conseils" --> NSERC/CRSNG, SSHRC/CRSH/SSRCC */
                 regexprmatches("^co(?:(?:un(?:cil|sel))|(?:nseil(?:s)?))$", lower(middle))
                 and (
                     -- The "middle" at the beginning of the fullname.
                     (   regexprmatches("^recherche(?:s)?(?:(?:\s(?:g(?:e|é)nie|science(?:s)?)(?:\s(?:humaines|naturel(?:les)?)?)?(?:\sg(?:e|é)nie)?)?)?\scanada.*", lower(next))    -- The term "canada" is put as mandatory here, as we get false-positives.
-                        or
-                        regexprmatches("^social\ssciences\shumanities\sresearch\scanada.*", lower(next))  -- It has the word "canada", so it's a canadian match.
+                        or regexprmatches("^social\sscience(?:s)?\shumanities\sresearch\scanada.*", lower(next))  -- It has the word "canada", so it's a canadian match.
                     )
                     or -- The "middle" at the end of the fullname.
                     (
-                        regexprmatches(".*(?:social|nat(?:ural|ional))\sscience(?:s)?\s(?:(?:engineering|humanities)(?:\sresearch)?|research)$", lower(prev))
-                        and -- Add this just to be more sure it's a "canadian" match..
-                        (
+                        ( regexprmatches(".*(?:social|nat(?:ural|ional))\sscience(?:s)?\s(?:(?:engineering|humanities)(?:\sresearch)?|research)$", lower(prev))
+                            or regexprmatches(".*humanities\sresearch$", lower(prev))
+                        )
+                        and (   -- Add this just to be more sure it's a "canadian" match..
                             regexprmatches("^canada.*", lower(next))
-                            or
-                            regexprmatches(".*canada\s(?:social|nat(?:ural|ional)).*", lower(prev))
+                            or regexprmatches(".*canada\s(?:social|nat(?:ural|ional)).*", lower(prev))
                         )
                     )
                 )
             )
             or (    /* Middle: "Canada", "Canadian" --> CIHR/IRSC */
                 regexprmatches("^canad(?:a|ian)$", lower(middle))   -- "Canadian" match for sure.
-                and (
-                    -- The "middle" at the beginning of the fullname
+                and ( -- The "middle" at the beginning of the fullname
                     regexprmatches("^institute(?:s)?\shealth\sresearch.*", lower(next))
-                    or
-                    -- The "middle" at the end of the fullname
+                    or -- The "middle" at the end of the fullname
                     regexprmatches(".*institut(?:(?:e)?(?:s)?)?\srecherche\ssant(?:e|é)$", lower(prev))
                 )
             )
         )
     )
     and (   /* Relation */
-        regexprmatches(".*(?:fund|support|financ|grant|sponsor|parrain|souten|subsidiz|promot|acquir|acknowledg|administ|assist|donor|bailleur|g(?:e|é)n(?:e|é)rosit).*", lower(prev||" "||next))
+        regexprmatches(".*(?:fund|support|financ|proje(?:c)?t|grant|subvention|sponsor|parrain|souten|subsidiz|promot|acquir|acknowledg|administ|assist|donor|bailleur|g(?:e|é)n(?:e|é)rosit).*", lower(prev||" "||next))
         or regexprmatches(".*(?:thank|gratefull|(?:re)?merci).*", lower(prev))
     )
 )
 where id is not null
-group by docid,id
+group by docid, id
 
 
 union all